About
We propose a methodological project which aims at combining individual-level data (from UK Biobank) with small area data (from administrative registries) to deal with the issue of unmeasured confounders that arises in small area epidemiological studies to assess the health impacts of environmental threats. We will investigate the adverse long-term physiological (hospital admission and mortality for cardio-respiratory diseases and cancer incidence / mortality) and psychiatric (depression and anxiety) effects associated with ambient air pollution, and noise from road and aircraft traffic. The project fits well with the UK Biobank?s mission of health related research in the public interest. We will integrate data from different sources, using innovative and rigorous statistical approaches, which will increase the quality and consequently the impact of public health research, leading to a better understanding of environmental determinants of cardio-respiratory diseases, cancers and psychiatric disorders at a population level. Outcomes from this project will include development of statistical methods to help improve inference combining different data sources and evidence to inform public health strategies to minimise adverse health effects from air pollution and noise exposures. We hypothesise that air and noise pollution each act as physiological stressors for the cardio-respiratory system and for some types of cancer and also can play a role in psychological distress. We will explore these hypotheses though the integration of different data sources (i.e. hospital episode statistics, vital statistics, high-resolution air and noise pollution estimates and UK Biobank data). We will look at single and multiple exposures to different air pollutants and noise. In particular, we will investigate whether the individual-level information from UK Biobank can help improve inference from the administrative data. We plan to conduct a national study, which gives good power to investigate environmental exposures with associated small excess risks, and to investigate variability at small area level. To ensure accurate results, we will need to consider the full UK Biobank cohort, which guarantee high spatial coverage (key aspect in small area studies). To carry out the proposed research we would need to have a spatial identifier for participants (e.g. Middle Super Output Area, see more details later) to be able to link the cohort with the administrative data which will also be aggregated on the same spatial units.